The following chapter contains the research design and methodology for this study.
The chapter also presents the research hypothesis that has been derived from the literature review and research questions. It explains the research procedure, sample, data collection method, measurement instrument, and data analysis methods. The chapter also provides details of validation and reliability.
Research Framework
For this study, the research framework was developed after conducting the literature review. In this research, we have four dependent variables: organizational trust, knowledge sharing, knowledge creation, and innovation. The independent variable for this research is business performance. In the next theoretical framework we show the relationship amongst these variables.
To be able to test the effects of the variables and their relations, this research developed the Organizational Trust, Knowledge Sharing, Knowledge Creation, Innovation and Business Performance by Cheng-Ping Shih and Ramos. This model is formed by partially adopting the model of Organizational Trust Model by Mooradian et al. (2006); Shockley-Zalabak et al. (2003), Knowledge Sharing Model by Cummings (2004); and Van den Hooff and De Ridder (2003), Knowledge Creation by Nonaka et al.
(1994), Innovation Model by Van der Panne et al. (2003), and Business Performance Model by Emden et al. (2005). The research framework developed can be found below.
H2
Figure 3.1. OK-OB model, developed by Cheng-Ping Shih and Fernando Ramos.
Organizational Trust
Research Hypotheses
The research hypotheses test the relationships among organizational trust, knowledge sharing, knowledge creation, innovation and business performance. Based on the literature review and research questions, the following null-hypotheses were developed for testing the results:
H1: Organizational trust has no effect on knowledge sharing.
H2: Organizational trust has no effect on knowledge creation.
H3: Knowledge creation has no effect on knowledge sharing.
H4: Knowledge sharing has no effect on innovation.
H5: Knowledge creation has no effect on innovation.
H6: Innovation has no effect on business performance.
H7: Organizational trust has no effect on business performance.
Research Procedure
In the next chart, we describe the procedure followed to complete the study.
Figure 3.2. Research procedures
1
• Research motivation
2
• Review literature
3
• Define constructs
4
• Identify research questions and hypotheses
5
• Develop research framework of the study
6
• Determine research methodology
7
• Develop instrument
8
• Conduct Pilot study
9
• Expert judgement and review
10
• Instrument review
11
• Data collection and coding
12
• Data analysis
13
• Conclusions and suggestions
Data Collection
This study used a quantitative approach, to its completion a self-reported questionnaire was used in order to gather all the data that was needed to test the effect of the variables in the study (organizational trust, knowledge sharing, and knowledge creation on innovation and business performance).
The participants selected for this research are Taiwanese employees from high technology companies. The respondents must be Taiwanese high technology company’s employees that are currently working in Taiwan. This means that employees could be from any region of Taiwan; as long as they are registered as employees in a high technology company when the research is conducted. Besides, the researcher used physical questionnaires to do the investigation. Participants have the privacy to fill in the questionnaire anonymously, so that they can provide honest answers. For this research the convenient sampling procedure was used to collect the data. The total population of the company is calculated at is calculated 2,100 employees by the end of February 2017. For the main study, a total of 220 questionnaires were collected from which 193 were valid questionnaires.
Measurement
For this study, we applied a quantitative approach, the objective of the research is to gather numerical data and then use it to analyze the statistical relationships amongst the variables. The instrument used for this research is a self-reported questionnaire. The function of the research instrument is to obtain the required data and then test the proposed hypotheses. The questionnaire divided the questions in six
sections according to their relevance and relationship. The questionnaire provides simple instructions to the respondents both at the beginning of the questionnaire and at the beginning of each section. The questions are applicable and answerable by most participants. All the constructs that are included in the questionnaire were adapted from pre-validated measures in existing correlated researches. This instrument consists of 5 variables with total of 67 questions. To briefly introduce the study for the respondents, a cover letter is used; it also explains its purpose. The survey consists of two parts: the first part consists of the questions about organizational trust, knowledge sharing, knowledge creation, innovation and business performance, and the second part is the respondent demographic data. For the first part, the participants were asked to rate each item in a scale ranging from “strongly disagree” (1) to “strongly agree” (5). For the second part, the participants were asked to choose one from the different options available. The measurements are described:
1. Organizational Trust (15 items): Adopted from Mooradian et al. (2006). the variables of trust consist of trust in peers (T_P) and trust in management (T_M) and by Shockley-Zalabak, Ellis and Cesaria (2003) the variables of trust consist of openness and honest (O_H), reliability (RE), identification (IDE). The items range from (1) “strongly disagree” to (5) “strongly agree”.
2. Knowledge Sharing (13 items): The measurement of knowledge sharing were adopted by Cummings’ (2004) scale and categorizing knowledge sharing into two types: intra-groups (I_G), and inter-groups (INTER_G) and by Van den Hooff and
De Ridder (2003) with the variables knowledge donating (K_D) and knowledge collecting (K_C). The items range from (1) “strongly disagree” to (5) “strongly agree”.
3. Knowledge Creation (11 items): Adopted from the questionnaire found in Knowledge Management Enablers, Processes, and Organizational Performance:
An Integrative View and Empirical Examination (Lee H. and Choi B 2013). The SECI model of Nonaka and Takeuchi is also used in this section to define the variables which are: Socialization (KC-S), Externalization (KC-E), Combination (KC-C) and Internalization (KC-I). The items range from (1) “strongly disagree” to (5) “strongly agree”.
4. Innovation (12 items): The questions were adopted from Van der Panne et al.
(2003). The questionnaire was modified and adapted for this research. The dimensions to be measured are Process Innovation (PI), Technological Innovation (TI) and Organizational Innovation (OI). The items range from (1) “strongly disagree” to (5) “strongly agree”.
5. Business Performance (12 items): The questions were adopted from Emden et al.
(2005); the questionnaire was modified and adapted for this research. The dimensions to be measured are Partnership Performance (PP), Market Performance (MP) and Financial Performance (FP). The items range from (1) “strongly
disagree” to (5) “strongly agree”.
The Respondent Demographic Profile: This part was added to provide a deeper analysis of the respondents, it contains demographic information about the participant’s gender, age, marital status, educational level and tenure.
Also, peer reviews and expert review were utilized to verify the validity of the instrument. Pilot test was initially conducted to ensure the validity of each item, before gathering the whole data for the study.
Construct Scales Coding
Before the analysis, the data was coded to facilitate the processing of information. The 67 items of the questionnaire were coded using a 5-point Likert scale as it was previously described in measurement. The items used to measure the variables of this study can be found below in Tables 3.1 to 3.6. Each item was coded for later use in the statistical analysis of the data by using PLS. For that purpose, the items were grouped within the same sub dimensions under the same construct. Thus, Dummy variables were created to code part II of the measurement instrument relating to demographics.
Table 3.1.
Items Measuring Organizational Trust
Construct Code Questionnaire Item
Trust in Peers (T_P; 3 items)
T_P1 If I got into difficulties at my company I know my co-workers would try and help me.
T_P2 I can trust the people I work with to lend me a hand if I meet the employee’s point of view.
T_M2 I feel quite confident that the company will always try to treat me fairly.
T_M3 Our working staff (managers, technicians, administrative, etc.) would be quite prepared to gain advantage by deceiving their co-workers.
Openness and Honesty (O_H;
3 items)
O_H1 I can get enough evaluation of my working abilities.
O_H2 I can have opinions to the decisions which are relevant to my work.
O_H3 When something is wrong, I am not afraid to tell my supervisor.
Reliability (RE; 3 items)
RE_1 My supervisor does what he or she says.
RE_2 Top managers keep their commitments to employees.
RE_3 My supervisor behaves in a consistent manner.
Identification (IDE; 3 items)
IDE_1 IDE_2 IDE_3
I feel connected to my peers.
I feel connected to my supervisors.
My value is similar to my co-worker’s values.
Table 3.2.
Items Measuring Knowledge Sharing
Construct Code Questionnaire Item
Intra-Groups (I_G; 3 items)
I_G1 I frequently share knowledge and information with my co-workers.
I_G2 I usually involve myself in discussions of various topics rather than specific topics with my co-workers.
I_G3 I usually spend a lot of time conducting knowledge sharing activities within my co-workers.
Inter-Groups (INTER_G; 3 items)
INTER_G1 I frequently share knowledge with co-workers even though he (or she) is not in my team.
INTER_G2 I usually involve myself in discussions of various topics rather than specific topics with workers from other co-worker’s teams.
INTER_G3 I spend a considerate amount of time conducting knowledge sharing activities with other co-worker teams in the company.
Knowledge Donating (K_D; 3 items)
K_D1 When I have learned something new, I tell my co-workers about it.
K_D2 When they have learned something new, my co-workers tell me about it.
K_D3 Knowledge sharing among co-workers is considered normal in my company.
K_C2 I share my skills with co-workers when they ask for it.
K_C3 Co-workers in my company share knowledge with me when I ask them to.
K_C4 Co-workers in my company share their skills with me when I ask them to.
Table 3.3.
Items Measuring Knowledge Creation
Construct Code Questionnaire Item
Socialization (KC_S; 3 items)
KC_S1 My company emphasizes sharing experience with other employees.
KC_S2 My company emphasizes finding new strategies and working opportunities by wandering inside the department.
KC_S3 My company emphasizes creating a working environment that allows peers to understand the technicalities of the
KC_E2 My company emphasizes creative and essential dialogues.
KC_E3 My company emphasizes the use of metaphors in dialogue for concept creation.
Combination (KC_C: 2 items)
KC_C1 My company emphasizes planning strategies by using published literature, computer simulation and forecasting.
KC_C2 My company emphasizes creating manuals and documents on products and services.
Internalization (KC_I; 3 items)
KC_I1 My company emphasizes forming teams, conducting experiments, and sharing results with the entire department.
KC_I2 My company emphasizes searching for and sharing new values and thoughts.
KC_I3 My company emphasizes bench-marking and test marketing.
Table 3.4.
Items Measuring Innovation
Construct Code Questionnaire Item
Process
Innovation (PI;
4 items)
PI1 My company enhances work efficiency because of company’s process innovation.
PI2 My company gets a higher success rate in launching new products and new services.
PI3
PI4
My company has clear and specific process of innovation development.
My company encourages using process innovation to understand the information of customers, suppliers and competitors.
OI1 My company improves internal communication efficiency because company’s organizational innovation.
OI2
OI3
OI4
My company cross-cultural communications ability is good for keeping ahead of market.
My company continues to import new way of management and knowledge to keep flexibility.
Employees highly accept the new way of innovational management.
Table 3.5.
Items Measuring Business Performance
Construct Code Questionnaire Item
Partnership Performance (PP; 4 items)
PP1 Comparing to our company’s competitors, company’s relationship with key alliance partners is strong.
PP2 Comparing to our company’s competitors, our company’s alliances are stable.
PP3
PP4
Comparing to our company’s competitors, company has an ability to sustain relationships regardless of changes in senior people.
Comparing to our company’s competitors, company and alliance partners keep commitments to each other.
Market Performance (TI; 4 items)
MP1 Comparing to our company’s competitors, company’s market development is good.
MP2 Comparing to our company’s competitors, company’s market share is high.
MP3
MP4
Comparing to our company’s competitors, company’s sales growth is satisfying.
Comparing to our company’s competitors, the products of company’s brand is popular.
Financial Performance (FP; 4 items)
FP1 Comparing to our company’s competitors, company’s profitability is satisfying.
FP2
FP3
FP4
Comparing to our company’s competitors, company’s return on investment is good.
Comparing to our company’s competitors, company’s Cost control is good.
Comparing to our company’s competitors, company’s Cash flow from operations is satisfying.
Table 3.6.
Items Measuring Demographic Data Profile
Variables Code Items
Gender 1 Male
2 Female
Age 1 21-25
2 26-30
3 31-35
4 36-40
5 41 and up
Educational Level 1 High School Diploma
2 Bachelor Degree
3 Master
4 PhD
Tenure 1 1-3
2 4-6
3 7-9
4 10-12
5 13 and up
Validity and Reliability
This test used the internal consistency and stability to test the validity and reliability. To test the internal consistency, the study used the Cronbach’s alpha values.
Cronbach’s alpha assessment is a common method that is used to test the reliability.
Accoding to Arambewela, Hall and Zuhair (2006) the minimum level for Cronbach’s alpha to be accepted is 7.0. As we can observe in the 4.1 table, this requirement was met.
In the work of Nunnally and Bernstein (1994) it was determined that the composite reliability needs to be higher or at least equal to 0.7, as we observed from the results of the study, the composite reliability also met that criteria. The average variance extracted (AVE) should at least be 0.5 or higher, this criteria was also met by the data gather in the study (Hair, Tatham and Anderson, 2006). As we can observe in the 4.1 table, the results show that the data had an AVE higher than 0.5, so that we can conclude that the data can be accepted. Furthermore, to be able to test the reliability of each individual item, we loadings were also examined. Chin (1998) determined that if the loadings are higher than 0.5, those can be accepted. This criterion was also met. If After observing the results from the Partial Least Square test, we can observe that the data collected for this study meets the minimum requirements of validity and reliability, then we conclude that is acceptable.
Partial Least Square (PLS)
Once the data is gathered for the study from the sample population, it was analyzed using Partial Least Square. Partial Least Square which is used to estimate the path models through the use of the latent variables. This procedure permits modeling
simultaneously the relation between the multiple constructs, or allows the analysis of a system of constructs. The main goal in PLS it is “maximize the explanation variance, thus R2 and the significance of relationships among constructs are measures indicative of how well a model is performing” (Bontis, 1998, p. 69). The Partial Least Square model is composed by two different parts, the first one it is the structural part, this part shows the relation amongst a measurement component and the variables, this shows what the relation between the variables is. To perform a confirmatory path and factor analysis the SmartPLS 2.0 software was used. To analyze the structural we evaluated the following factors: The path coefficients, the coefficient of determination (R2), bootstrapping and t-value (t). The function of the R square is to explain the endogenous latent variables in relation to the total variance. In his study, Chin (1998) defined the R2 values of significance to be .67 (considered substantial), .33 (considered moderate) and .19 (considered weak). The function of the path coefficients is to judge the relation that exist among the variables, the also help to determine what is the relation’s direction and how significant it is. In Partial Least Squares this values can be observed by bootstrapping the collected data to later examine the t-value results. Hair, Ringle, and Sarstedt (2011) has previously defined what the critical values are in a two-tailed test, this tests has determined that the results are weak if they are lower than 1.65, moderate when they are between 1.65 and strong if they are equal or higher than 1.96 2.58.
Bootstrapping can be defined as “a nonparametric approach to statistical inference that does not make any distributional assumptions of the parameters like traditional methods. It draws conclusions of a population strictly from the sample at hand (Sharma
& Kim, 2012). Wong (2010) said that Partial Least Squares can be considered as a good
alternative under the following circumstances; when there is little available theory for the applications, when we cannot ensure the correct model specification and predictive accuracy is paramount and finally when the size sample is considered small. Partial Least Square was considered to be ideal for this research for the reasons presented as follows:
1. This research is exploratory rather than confirmatory in nature.
2. Partial Least Squares have less strict regulations regarding levels of correlation between variables, sample size, and distribution parameters.
3. Partial Least Squares is able test the psychometric properties of the indices, and hence it can gives good evidence about the relations within variables.